The problem

Sidewalk delivery is cost-constrained, not demand-constrained. Demand for last-mile is large and structurally growing — the blocker is that per-delivery rates erase the thin margins of restaurants, retailers, and logistics operators.

The solution: engineer the robot from the cost backward

Frugal robot. Mass-produced SoC and sensors instead of expensive automotive stack. AI trained entirely in a custom simulator, deployed on real streets without fine-tuning. Hardware can be assembled by entry-level technicians, enabling decentralized manufacturing. Demo.

Platform model. Manvel Robotics develops tech, runs platform and order distribution; regional partners operate fleets. Asset-light, easy to scale on volume + geography. Deployment is partner-driven and city-by-city, into geographies where the cost case is strongest and customer pickup friction is lowest.

Why now

  • Labor scarcity is structural, not cyclical — aging populations and courier shortages make automated last-mile a necessity
  • Commodity SoCs with NPUs now support real-time perception & AI
  • Funded competitors haven’t cracked unit economics — gap is open for cost-driven entrants

Traction

The Deliverator prototype
  • Working prototype operational on real urban sidewalks
  • Campus pilot launching summer 2026

Team

Manvel Avetisian — Founder. 20+ years in AI/ML at Google, Yandex, Sber AI Lab. Led 100+ researchers. Previously Co-Founder at AIvarix (healthcare AI, sold to Servier in 2025). Published researcher, h-index 9.

Market

  • Sidewalk delivery is projected at $3.2B by 2030 (32% CAGR) — but that figure only counts orders viable at today’s costs.
  • Our cost advantage creates a new market below that floor: delivery becomes economic for price-conscious customers — thin-margin restaurants, low-AOV retailers, logistics operators — priced out today. We create demand, not just capture share.

Contact